Why a table in a PDF should stay a table in Excel
Getting a table as a PDF, from a bank statement, a report or a price list, is the easy part. It gets harder once the numbers need further work: totalled, filtered, pulled into a spreadsheet of your own. For that, rows and columns need to become an actual Excel table with separate cells again.
Convert PDF to Excel detects the table structure automatically and places each column in its own cell, instead of pulling in the whole text as one block.
What tends to go wrong during detection
Tables without visible lines cause the most trouble, because the software then has to judge where a column ends purely by spacing. With tightly packed numbers, that can merge two columns into one. A quick look at the first few rows after conversion shows right away whether that happened.
Scanned tables are a different case
If the PDF is actually a photo or a scan behind the scenes, there is no recognisable text yet at all, only pixels. In that case text recognition needs to run first. PDF text recognition makes the content searchable, and only then does the table convert sensibly into Excel.
Several tables in one document
If several tables sit in the same file, across different pages of an annual report for example, each one lands on its own sheet. That saves the manual work of figuring out which rows belong together.
What still needs checking afterwards
Numbers with thousand separators or currency symbols sometimes get pulled in as text instead of as numbers, which blocks later calculations. A quick test with a simple sum formula shows whether Excel treats the column as a number or not.
When only a few cells are needed
If only one table out of a longer report matters, Extract PDF pages saves time beforehand: pull out the right page first, then convert it, instead of processing the whole document.
Common questions
Do formulas from the PDF survive? No, a PDF only stores the result, not a formula. After conversion, fixed values remain.
Does this work with several tables per page? Yes, they get detected individually as long as there is a visible gap between them.
What happens to text outside the table? It gets carried over too, usually outside the recognised table area.